KNOWLEDGE-BASED LEARNING IN EXPLORATORY SCIENCE - LEARNING RULES TO PREDICT RODENT CARCINOGENICITY

In this paper, we report on a multi-year collaboration among computer scientists, toxicologists, chemists, and a statistician, in which the RL induction program was used to assist toxicologists in analyzing rel ationships among various features of chemical compounds and their carc inogenicity in rodents. Our investigation demonstrated the utility of knowledge-based rule induction in the problem of predicting rodent car cinogenicity and the place of rule induction in the overall process of discovery. Flexibility of the program in accepting different definiti ons of background knowledge and preferences was considered essential i n this exploratory effort. This investigation has made significant con tributions not only to predicting carcinogenicity and non-carcinogenic ity in rodents, but to understanding how to extend a rule induction pr ogram into an exploratory data analysis tool.